Open main menu
Home
Random
Recent changes
Special pages
Community portal
Preferences
About Wikipedia
Disclaimers
Incubator escapee wiki
Search
User menu
Talk
Dark mode
Contributions
Create account
Log in
Editing
Artificial intelligence
(section)
Warning:
You are not logged in. Your IP address will be publicly visible if you make any edits. If you
log in
or
create an account
, your edits will be attributed to your username, along with other benefits.
Anti-spam check. Do
not
fill this in!
==Ethics== {{Main|Ethics of artificial intelligence}} AI has potential benefits and potential risks.<ref>{{Cite web |title=Ethics of Artificial Intelligence and Robotics |url=https://plato.stanford.edu/archives/fall2023/entries/ethics-ai/ |website=Stanford Encyclopedia of Philosophy Archive |date=30 April 2020 |last1=Müller |first1=Vincent C. |access-date=5 October 2024 |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005165650/https://plato.stanford.edu/archives/fall2023/entries/ethics-ai/ |url-status=live }}</ref> AI may be able to advance science and find solutions for serious problems: [[Demis Hassabis]] of [[DeepMind]] hopes to "solve intelligence, and then use that to solve everything else".{{Sfnp|Simonite|2016}} However, as the use of AI has become widespread, several unintended consequences and risks have been identified.{{Sfnp|Russell|Norvig|2021|p=987}} In-production systems can sometimes not factor ethics and bias into their AI training processes, especially when the AI algorithms are inherently unexplainable in deep learning.{{Sfnp|Laskowski|2023}} === Risks and harm === ==== Privacy and copyright ==== {{Further|Information privacy|Artificial intelligence and copyright}} Machine learning algorithms require large amounts of data. The techniques used to acquire this data have raised concerns about [[privacy]], [[surveillance]] and [[copyright]]. <!-- PRIVACY PROBLEM --> AI-powered devices and services, such as virtual assistants and IoT products, continuously collect personal information, raising concerns about intrusive data gathering and unauthorized access by third parties. The loss of privacy is further exacerbated by AI's ability to process and combine vast amounts of data, potentially leading to a surveillance society where individual activities are constantly monitored and analyzed without adequate safeguards or transparency. Sensitive user data collected may include online activity records, geolocation data, video, or audio.{{Sfnp|GAO|2022}} For example, in order to build [[speech recognition]] algorithms, [[Amazon (company)|Amazon]] has recorded millions of private conversations and allowed [[temporary worker]]s to listen to and transcribe some of them.{{Sfnp|Valinsky|2019}} Opinions about this widespread surveillance range from those who see it as a [[necessary evil]] to those for whom it is clearly [[unethical]] and a violation of the [[right to privacy]].{{Sfnp|Russell|Norvig|2021|p=991}} <!-- PRIVACY SOLUTIONS --> AI developers argue that this is the only way to deliver valuable applications and have developed several techniques that attempt to preserve privacy while still obtaining the data, such as [[data aggregation]], [[de-identification]] and [[differential privacy]].{{Sfnp|Russell|Norvig|2021|pp=991–992}} Since 2016, some privacy experts, such as [[Cynthia Dwork]], have begun to view privacy in terms of [[fairness (machine learning)|fairness]]. [[Brian Christian]] wrote that experts have pivoted "from the question of 'what they know' to the question of 'what they're doing with it'."{{Sfnp|Christian|2020|p=63}} <!-- COPYRIGHT AND GENERATIVE AI --> Generative AI is often trained on unlicensed copyrighted works, including in domains such as images or computer code; the output is then used under the rationale of "[[fair use]]". Experts disagree about how well and under what circumstances this rationale will hold up in courts of law; relevant factors may include "the purpose and character of the use of the copyrighted work" and "the effect upon the potential market for the copyrighted work".{{Sfnp|Vincent|2022}}<ref>{{Cite web |last=Kopel |first=Matthew |title=Copyright Services: Fair Use |url=https://guides.library.cornell.edu/copyright/fair-use |access-date=2024-04-26 |website=Cornell University Library |archive-date=26 September 2024 |archive-url=https://web.archive.org/web/20240926194057/https://guides.library.cornell.edu/copyright/fair-use |url-status=live }}</ref> Website owners who do not wish to have their content scraped can indicate it in a "[[robots.txt]]" file.<ref>{{Cite magazine |last=Burgess |first=Matt |title=How to Stop Your Data From Being Used to Train AI |url=https://www.wired.com/story/how-to-stop-your-data-from-being-used-to-train-ai |access-date=2024-04-26 |magazine=Wired |issn=1059-1028 |archive-date=3 October 2024 |archive-url=https://web.archive.org/web/20241003180100/https://www.wired.com/story/how-to-stop-your-data-from-being-used-to-train-ai/ |url-status=live }}</ref> In 2023, leading authors (including [[John Grisham]] and [[Jonathan Franzen]]) sued AI companies for using their work to train generative AI.{{Sfnp|Reisner|2023}}{{Sfnp|Alter|Harris|2023}} Another discussed approach is to envision a separate ''[[sui generis]]'' system of protection for creations generated by AI to ensure fair attribution and compensation for human authors.<ref>{{Cite web |title=Getting the Innovation Ecosystem Ready for AI. An IP policy toolkit |url=https://www.wipo.int/edocs/pubdocs/en/wipo-pub-2003-en-getting-the-innovation-ecosystem-ready-for-ai.pdf |website=[[WIPO]]}}</ref> ====Dominance by tech giants==== The commercial AI scene is dominated by [[Big Tech]] companies such as [[Alphabet Inc.]], [[Amazon (company)|Amazon]], [[Apple Inc.]], [[Meta Platforms]], and [[Microsoft]].<ref>{{Cite web |last=Hammond |first=George |date=27 December 2023 |title=Big Tech is spending more than VC firms on AI startups |url=https://arstechnica.com/ai/2023/12/big-tech-is-spending-more-than-vc-firms-on-ai-startups |url-status=live |archive-url=https://web.archive.org/web/20240110195706/https://arstechnica.com/ai/2023/12/big-tech-is-spending-more-than-vc-firms-on-ai-startups |archive-date=Jan 10, 2024 |website=Ars Technica}}</ref><ref>{{Cite web |last=Wong |first=Matteo |date=24 October 2023 |title=The Future of AI Is GOMA |url=https://www.theatlantic.com/technology/archive/2023/10/big-ai-silicon-valley-dominance/675752 |url-access=subscription |url-status=live |archive-url=https://web.archive.org/web/20240105020744/https://www.theatlantic.com/technology/archive/2023/10/big-ai-silicon-valley-dominance/675752 |archive-date=Jan 5, 2024 |website=The Atlantic |ref=none}}</ref><ref>{{Cite news |date=Mar 26, 2023 |title=Big tech and the pursuit of AI dominance |url=https://www.economist.com/business/2023/03/26/big-tech-and-the-pursuit-of-ai-dominance |url-access=subscription |url-status=live |archive-url=https://web.archive.org/web/20231229021351/https://www.economist.com/business/2023/03/26/big-tech-and-the-pursuit-of-ai-dominance |archive-date=Dec 29, 2023 |newspaper=The Economist}}</ref> Some of these players already own the vast majority of existing [[cloud computing|cloud infrastructure]] and [[computing]] power from [[data center]]s, allowing them to entrench further in the marketplace.<ref>{{Cite news |last=Fung |first=Brian |date=19 December 2023 |title=Where the battle to dominate AI may be won |url=https://www.cnn.com/2023/12/19/tech/cloud-competition-and-ai/index.html |url-status=live |archive-url=https://web.archive.org/web/20240113053332/https://www.cnn.com/2023/12/19/tech/cloud-competition-and-ai/index.html |archive-date=Jan 13, 2024 |work=CNN Business}}</ref><ref>{{Cite news |last=Metz |first=Cade |date=5 July 2023 |title=In the Age of A.I., Tech's Little Guys Need Big Friends |url=https://www.nytimes.com/2023/07/05/business/artificial-intelligence-power-data-centers.html |work=The New York Times |access-date=5 October 2024 |archive-date=8 July 2024 |archive-url=https://web.archive.org/web/20240708214644/https://www.nytimes.com/2023/07/05/business/artificial-intelligence-power-data-centers.html |url-status=live }}</ref> ====Power needs and environmental impacts==== {{See also|Environmental impacts of artificial intelligence}} In January 2024, the [[International Energy Agency]] (IEA) released ''Electricity 2024, Analysis and Forecast to 2026'', forecasting electric power use.<ref>{{Cite web |date=2024-01-24 |title=Electricity 2024 – Analysis |url=https://www.iea.org/reports/electricity-2024 |access-date=2024-07-13 |website=IEA}}</ref> This is the first IEA report to make projections for data centers and power consumption for artificial intelligence and cryptocurrency. The report states that power demand for these uses might double by 2026, with additional electric power usage equal to electricity used by the whole Japanese nation.<ref>{{Cite web |last=Calvert |first=Brian |date=28 March 2024 |title=AI already uses as much energy as a small country. It's only the beginning. |url=https://www.vox.com/climate/2024/3/28/24111721/ai-uses-a-lot-of-energy-experts-expect-it-to-double-in-just-a-few-years |website=Vox |location=New York, New York |access-date=5 October 2024 |archive-date=3 July 2024 |archive-url=https://web.archive.org/web/20240703080555/https://www.vox.com/climate/2024/3/28/24111721/ai-uses-a-lot-of-energy-experts-expect-it-to-double-in-just-a-few-years |url-status=live }}</ref> Prodigious power consumption by AI is responsible for the growth of fossil fuels use, and might delay closings of obsolete, carbon-emitting coal energy facilities. There is a feverish rise in the construction of data centers throughout the US, making large technology firms (e.g., Microsoft, Meta, Google, Amazon) into voracious consumers of electric power. Projected electric consumption is so immense that there is concern that it will be fulfilled no matter the source. A ChatGPT search involves the use of 10 times the electrical energy as a Google search. The large firms are in haste to find power sources – from nuclear energy to geothermal to fusion. The tech firms argue that – in the long view – AI will be eventually kinder to the environment, but they need the energy now. AI makes the power grid more efficient and "intelligent", will assist in the growth of nuclear power, and track overall carbon emissions, according to technology firms.<ref>{{Cite news |last1=Halper |first1=Evan |last2=O'Donovan |first2=Caroline |date=21 June 2024 |title=AI is exhausting the power grid. Tech firms are seeking a miracle solution. |url=https://www.washingtonpost.com/business/2024/06/21/artificial-intelligence-nuclear-fusion-climate/?utm_campaign=wp_post_most&utm_medium=email&utm_source=newsletter&wpisrc=nl_most&carta-url=https%3A%2F%2Fs2.washingtonpost.com%2Fcar-ln-tr%2F3e0d678%2F6675a2d2c2c05472dd9ec0f4%2F596c09009bbc0f20865036e7%2F12%2F52%2F6675a2d2c2c05472dd9ec0f4 |newspaper=Washington Post}}</ref> A 2024 [[Goldman Sachs]] Research Paper, ''AI Data Centers and the Coming US Power Demand Surge'', found "US power demand (is) likely to experience growth not seen in a generation...." and forecasts that, by 2030, US data centers will consume 8% of US power, as opposed to 3% in 2022, presaging growth for the electrical power generation industry by a variety of means.<ref>{{Cite web |last=Davenport |first=Carly |title=AI Data Centers and the Coming YS Power Demand Surge |url=https://www.goldmansachs.com/intelligence/pages/gs-research/generational-growth-ai-data-centers-and-the-coming-us-power-surge/report.pdf |website=Goldman Sachs |access-date=5 October 2024 |archive-date=26 July 2024 |archive-url=https://web.archive.org/web/20240726080428/https://www.goldmansachs.com/intelligence/pages/gs-research/generational-growth-ai-data-centers-and-the-coming-us-power-surge/report.pdf |url-status=dead }}</ref> Data centers' need for more and more electrical power is such that they might max out the electrical grid. The Big Tech companies counter that AI can be used to maximize the utilization of the grid by all.<ref>{{Cite news |last=Ryan |first=Carol |date=12 April 2024 |title=Energy-Guzzling AI Is Also the Future of Energy Savings |url=https://www.wsj.com/business/energy-oil/ai-data-centers-energy-savings-d602296e |work=Wall Street Journal |publisher=Dow Jones}}</ref> In 2024, the ''Wall Street Journal'' reported that big AI companies have begun negotiations with the US nuclear power providers to provide electricity to the data centers. In March 2024 Amazon purchased a Pennsylvania nuclear-powered data center for $650 Million (US).<ref>{{Cite news |last=Hiller |first=Jennifer |date=1 July 2024 |title=Tech Industry Wants to Lock Up Nuclear Power for AI |url=https://www.wsj.com/business/energy-oil/tech-industry-wants-to-lock-up-nuclear-power-for-ai-6cb75316?mod=djem10point |work=Wall Street Journal |publisher=Dow Jones |access-date=5 October 2024 |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005165650/https://www.wsj.com/business/energy-oil/tech-industry-wants-to-lock-up-nuclear-power-for-ai-6cb75316?mod=djem10point |url-status=live }}</ref> [[Nvidia]] CEO [[Jen-Hsun Huang]] said nuclear power is a good option for the data centers.<ref>{{Cite news |last1=Kendall |first1=Tyler |date=28 September 2024 |title=Nvidia's Huang Says Nuclear Power an Option to Feed Data Centers |url=https://www.bloomberg.com/news/articles/2024-09-27/nvidia-s-huang-says-nuclear-power-an-option-to-feed-data-centers |newspaper=Bloomberg}}</ref> In September 2024, [[Microsoft]] announced an agreement with [[Constellation Energy]] to re-open the [[Three Mile Island]] nuclear power plant to provide Microsoft with 100% of all electric power produced by the plant for 20 years. Reopening the plant, which suffered a partial nuclear meltdown of its Unit 2 reactor in 1979, will require Constellation to get through strict regulatory processes which will include extensive safety scrutiny from the US [[Nuclear Regulatory Commission]]. If approved (this will be the first ever US re-commissioning of a nuclear plant), over 835 megawatts of power – enough for 800,000 homes – of energy will be produced. The cost for re-opening and upgrading is estimated at $1.6 billion (US) and is dependent on tax breaks for nuclear power contained in the 2022 US [[Inflation Reduction Act]].<ref>{{Cite news |last=Halper |first=Evan |date=20 September 2024 |title=Microsoft deal would reopen Three Mile Island nuclear plant to power AI |url=https://www.washingtonpost.com/business/2024/09/20/microsoft-three-mile-island-nuclear-constellation |newspaper=Washington Post}}</ref> The US government and the state of Michigan are investing almost $2 billion (US) to reopen the [[Palisades Nuclear Generating Station|Palisades Nuclear]] reactor on Lake Michigan. Closed since 2022, the plant is planned to be reopened in October 2025. The Three Mile Island facility will be renamed the Crane Clean Energy Center after Chris Crane, a nuclear proponent and former CEO of [[Exelon]] who was responsible for Exelon spinoff of Constellation.<ref>{{Cite news |last=Hiller |first=Jennifer |date=20 September 2024 |title=Three Mile Island's Nuclear Plant to Reopen, Help Power Microsoft's AI Centers |url=https://www.wsj.com/business/energy-oil/three-mile-islands-nuclear-plant-to-reopen-help-power-microsofts-ai-centers-aebfb3c8?mod=Searchresults_pos1&page=1 |work=Wall Street Journal |publisher=Dow Jones |access-date=5 October 2024 |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170152/https://www.wsj.com/business/energy-oil/three-mile-islands-nuclear-plant-to-reopen-help-power-microsofts-ai-centers-aebfb3c8?mod=Searchresults_pos1&page=1 |url-status=live }}</ref> After the last approval in September 2023, [[Taiwan]] suspended the approval of data centers north of [[Taoyuan, Taiwan|Taoyuan]] with a capacity of more than 5 MW in 2024, due to power supply shortages.<ref name="DatacenterDynamics">{{Cite news |author=Niva Yadav |date=19 August 2024 |title=Taiwan to stop large data centers in the North, cites insufficient power |url=https://www.datacenterdynamics.com/en/news/taiwan-to-stop-large-data-centers-in-the-north-cites-insufficient-power/ |publisher=DatacenterDynamics |archive-date=8 November 2024 |access-date=7 November 2024 |archive-url=https://web.archive.org/web/20241108213650/https://www.datacenterdynamics.com/en/news/taiwan-to-stop-large-data-centers-in-the-north-cites-insufficient-power/ |url-status=live }}</ref> Taiwan aims to [[Nuclear power phase-out|phase out nuclear power]] by 2025.<ref name="DatacenterDynamics" /> On the other hand, [[Singapore]] imposed a ban on the opening of data centers in 2019 due to electric power, but in 2022, lifted this ban.<ref name="DatacenterDynamics" /> Although most nuclear plants in Japan have been shut down after the 2011 [[Fukushima nuclear accident]], according to an October 2024 ''Bloomberg'' article in Japanese, cloud gaming services company Ubitus, in which Nvidia has a stake, is looking for land in Japan near nuclear power plant for a new data center for generative AI.<ref name=bloombergjp>{{Cite news |last1=Mochizuki |first1=Takashi |last2=Oda |first2=Shoko |date=18 October 2024 |title=エヌビディア出資の日本企業、原発近くでAIデータセンター新設検討 |url=https://www.bloomberg.co.jp/news/articles/2024-10-18/SLHGKKT0AFB400 |newspaper=Bloomberg |language=Japanese |archive-date=8 November 2024 |access-date=7 November 2024 |archive-url=https://web.archive.org/web/20241108213843/https://www.bloomberg.co.jp/news/articles/2024-10-18/SLHGKKT0AFB400 |url-status=live }}</ref> Ubitus CEO Wesley Kuo said nuclear power plants are the most efficient, cheap and stable power for AI.<ref name=bloombergjp /> On 1 November 2024, the [[Federal Energy Regulatory Commission]] (FERC) rejected an application submitted by [[Talen Energy]] for approval to supply some electricity from the nuclear power station [[Susquehanna Steam Electric Station|Susquehanna]] to Amazon's data center.<ref name="Bloomberg20241104">{{Cite news |author=Naureen S Malik and Will Wade |date=5 November 2024 |title=Nuclear-Hungry AI Campuses Need New Plan to Find Power Fast |url=https://www.bloomberg.com/news/articles/2024-11-04/nuclear-hungry-ai-campuses-need-new-strategy-to-find-power-fast |publisher=Bloomberg}}</ref> According to the Commission Chairman [[Willie L. Phillips]], it is a burden on the electricity grid as well as a significant cost shifting concern to households and other business sectors.<ref name="Bloomberg20241104" /> In 2025 a report prepared by the International Energy Agency estimated the [[greenhouse gas emissions]] from the energy consumption of AI at 180 million tons. By 2035, these emissions could rise to 300-500 million tonnes depending on what measures will be taken. This is below 1.5% of the energy sector emissions. The emissions reduction potential of AI was estimated at 5% of the energy sector emissions, but [[Rebound effect (conservation)|rebound effects]] (for example if people will pass from public transport to autonomous cars) can reduce it.<ref>{{cite web |title=Energy and AI Executive summary |url=https://www.iea.org/reports/energy-and-ai/executive-summary |website=International Energy Agency |access-date=10 April 2025}}</ref> ==== Misinformation ==== {{See also|YouTube#Moderation and offensive content}} [[YouTube]], [[Facebook]] and others use [[recommender system]]s to guide users to more content. These AI programs were given the goal of [[mathematical optimization|maximizing]] user engagement (that is, the only goal was to keep people watching). The AI learned that users tended to choose [[misinformation]], [[conspiracy theories]], and extreme [[partisan (politics)|partisan]] content, and, to keep them watching, the AI recommended more of it. Users also tended to watch more content on the same subject, so the AI led people into [[filter bubbles]] where they received multiple versions of the same misinformation.{{Sfnp|Nicas|2018}} This convinced many users that the misinformation was true, and ultimately undermined trust in institutions, the media and the government.<ref>{{Cite web |last1=Rainie |first1=Lee |last2=Keeter |first2=Scott |last3=Perrin |first3=Andrew |date=July 22, 2019 |title=Trust and Distrust in America |url=https://www.pewresearch.org/politics/2019/07/22/trust-and-distrust-in-america |url-status=live |archive-url=https://web.archive.org/web/20240222000601/https://www.pewresearch.org/politics/2019/07/22/trust-and-distrust-in-america |archive-date=Feb 22, 2024 |website=Pew Research Center}}</ref> The AI program had correctly learned to maximize its goal, but the result was harmful to society. After the U.S. election in 2016, major technology companies took some steps to mitigate the problem.<ref>{{Cite magazine |last=Kosoff |first=Maya |date=2018-02-08 |title=YouTube Struggles to Contain Its Conspiracy Problem |url=https://www.vanityfair.com/news/2018/02/youtube-conspiracy-problem |access-date=2025-04-10 |magazine=Vanity Fair |language=en-US}}</ref> In 2022, [[generative AI]] began to create images, audio, video and text that are indistinguishable from real photographs, recordings, films, or human writing. It is possible for bad actors to use this technology to create massive amounts of misinformation or propaganda.{{Sfnp|Williams|2023}} One such potential malicious use is deepfakes for [[computational propaganda]].<ref>{{Cite journal |last=Olanipekun |first=Samson Olufemi |date=2025 |title=Computational propaganda and misinformation: AI technologies as tools of media manipulation |url=https://journalwjarr.com/node/366 |journal=World Journal of Advanced Research and Reviews |language=en |volume=25 |issue=1 |pages=911–923 |doi=10.30574/wjarr.2025.25.1.0131 |issn=2581-9615}}</ref> AI pioneer [[Geoffrey Hinton]] expressed concern about AI enabling "authoritarian leaders to manipulate their electorates" on a large scale, among other risks.{{Sfnp|Taylor|Hern|2023}} AI researchers at [[Microsoft]], [[OpenAI]], universities and other organisations have suggested using "[[Proof of personhood#Approaches|personhood credentials]]" as a way to overcome online deception enabled by AI models.<ref>{{Cite news |title=To fight AI, we need 'personhood credentials,' say AI firms |url=https://www.theregister.com/2024/09/03/ai_personhood_credentials/ |archive-url=http://web.archive.org/web/20250424232537/https://www.theregister.com/2024/09/03/ai_personhood_credentials/ |archive-date=2025-04-24 |access-date=2025-05-09 |language=en}}</ref> ====Algorithmic bias and fairness==== {{Main|Algorithmic bias|Fairness (machine learning)}} Machine learning applications will be [[algorithmic bias|biased]]{{Efn|In statistics, a [[Bias (statistics)|bias]] is a systematic error or deviation from the correct value. But in the context of [[Fairness (machine learning)|fairness]], it refers to a tendency in favor or against a certain group or individual characteristic, usually in a way that is considered unfair or harmful. A statistically unbiased AI system that produces disparate outcomes for different demographic groups may thus be viewed as biased in the ethical sense.<ref name="Samuel-2022"/>}} if they learn from biased data.{{Sfnp|Rose|2023}} The developers may not be aware that the bias exists.{{Sfnp|CNA|2019}} Bias can be introduced by the way [[training data]] is selected and by the way a model is deployed.{{Sfnp|Goffrey|2008|p=17}}{{Sfnp|Rose|2023}} If a biased algorithm is used to make decisions that can seriously [[harm]] people (as it can in [[health equity|medicine]], [[credit rating|finance]], [[recruitment]], [[public housing|housing]] or [[policing]]) then the algorithm may cause [[discrimination]].<ref>{{Harvtxt|Berdahl|Baker|Mann|Osoba|2023}}; {{Harvtxt|Goffrey|2008|p=17}}; {{Harvtxt|Rose|2023}}; {{Harvtxt|Russell|Norvig|2021|p=995}}</ref> The field of [[fairness (machine learning)|fairness]] studies how to prevent harms from algorithmic biases. On June 28, 2015, [[Google Photos]]'s new image labeling feature mistakenly identified Jacky Alcine and a friend as "gorillas" because they were black. The system was trained on a dataset that contained very few images of black people,{{Sfnp|Christian|2020|p=25}} a problem called "sample size disparity".{{Sfnp|Russell|Norvig|2021|p=995}} Google "fixed" this problem by preventing the system from labelling ''anything'' as a "gorilla". Eight years later, in 2023, Google Photos still could not identify a gorilla, and neither could similar products from Apple, Facebook, Microsoft and Amazon.{{Sfnp|Grant|Hill|2023}} [[COMPAS (software)|COMPAS]] is a commercial program widely used by [[U.S. court]]s to assess the likelihood of a [[defendant]] becoming a [[recidivist]]. In 2016, [[Julia Angwin]] at [[ProPublica]] discovered that COMPAS exhibited racial bias, despite the fact that the program was not told the races of the defendants. Although the error rate for both whites and blacks was calibrated equal at exactly 61%, the errors for each race were different—the system consistently overestimated the chance that a black person would re-offend and would underestimate the chance that a white person would not re-offend.{{Sfnp|Larson|Angwin|2016}} In 2017, several researchers{{Efn|Including [[Jon Kleinberg]] ([[Cornell University]]), Sendhil Mullainathan ([[University of Chicago]]), Cynthia Chouldechova ([[Carnegie Mellon]]) and Sam Corbett-Davis ([[Stanford]]){{Sfnp|Christian|2020|p=67–70}}}} showed that it was mathematically impossible for COMPAS to accommodate all possible measures of fairness when the base rates of re-offense were different for whites and blacks in the data.<ref>{{Harvtxt|Christian|2020|pp=67–70}}; {{Harvtxt|Russell|Norvig|2021|pp=993–994}}</ref> A program can make biased decisions even if the data does not explicitly mention a problematic feature (such as "race" or "gender"). The feature will correlate with other features (like "address", "shopping history" or "first name"), and the program will make the same decisions based on these features as it would on "race" or "gender".<ref>{{Harvtxt|Russell|Norvig|2021|p=995}}; {{Harvtxt|Lipartito|2011|p=36}}; {{Harvtxt|Goodman|Flaxman|2017|p=6}}; {{Harvtxt|Christian|2020|pp=39–40, 65}}</ref> Moritz Hardt said "the most robust fact in this research area is that fairness through blindness doesn't work."<ref>Quoted in {{Harvtxt|Christian|2020|p=65}}.</ref> Criticism of COMPAS highlighted that machine learning models are designed to make "predictions" that are only valid if we assume that the future will resemble the past. If they are trained on data that includes the results of racist decisions in the past, machine learning models must predict that racist decisions will be made in the future. If an application then uses these predictions as ''recommendations'', some of these "recommendations" will likely be racist.<ref>{{Harvtxt|Russell|Norvig|2021|p=994}}; {{Harvtxt|Christian|2020|pp=40, 80–81}}</ref> Thus, machine learning is not well suited to help make decisions in areas where there is hope that the future will be ''better'' than the past. It is descriptive rather than prescriptive.{{Efn|Moritz Hardt (a director at the [[Max Planck Institute for Intelligent Systems]]) argues that machine learning "is fundamentally the wrong tool for a lot of domains, where you're trying to design interventions and mechanisms that change the world."<ref>Quoted in {{Harvtxt|Christian|2020|p=80}}</ref>}} Bias and unfairness may go undetected because the developers are overwhelmingly white and male: among AI engineers, about 4% are black and 20% are women.{{Sfnp|Russell|Norvig|2021|p=995}} There are various conflicting definitions and mathematical models of fairness. These notions depend on ethical assumptions, and are influenced by beliefs about society. One broad category is [[Distributive justice|distributive fairness]], which focuses on the outcomes, often identifying groups and seeking to compensate for statistical disparities. Representational fairness tries to ensure that AI systems do not reinforce negative [[stereotype]]s or render certain groups invisible. Procedural fairness focuses on the decision process rather than the outcome. The most relevant notions of fairness may depend on the context, notably the type of AI application and the stakeholders. The subjectivity in the notions of bias and fairness makes it difficult for companies to operationalize them. Having access to sensitive attributes such as race or gender is also considered by many AI ethicists to be necessary in order to compensate for biases, but it may conflict with [[anti-discrimination law]]s.<ref name="Samuel-2022">{{Cite web |last=Samuel |first=Sigal |date=2022-04-19 |title=Why it's so damn hard to make AI fair and unbiased |url=https://www.vox.com/future-perfect/22916602/ai-bias-fairness-tradeoffs-artificial-intelligence |access-date=2024-07-24 |website=Vox |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170153/https://www.vox.com/future-perfect/22916602/ai-bias-fairness-tradeoffs-artificial-intelligence |url-status=live }}</ref> At its 2022 [[ACM Conference on Fairness, Accountability, and Transparency|Conference on Fairness, Accountability, and Transparency]] (ACM FAccT 2022), the [[Association for Computing Machinery]], in Seoul, South Korea, presented and published findings that recommend that until AI and robotics systems are demonstrated to be free of bias mistakes, they are unsafe, and the use of self-learning neural networks trained on vast, unregulated sources of flawed internet data should be curtailed.{{Dubious|date=July 2024|reason=Depending on what is meant by "free of bias", it may be impossible in practice to demonstrate it. Additionally, the study evaluates the priors (initial assumptions) of the robots, rather than their decision-making in scenarios where there is a correct choice. For example, it may not be sexist to have the prior that most doctors are males (it's actually an accurate statistical prior in the world we currently live in, so the bias may arguably be to not have this prior). If forced to choose which one is the doctor based solely on gender, a rational person seeking to maximize the number of correct answers would choose the man 100% of the time. The real issue arises when such priors lead to significant discrimination.}}{{Sfnp|Dockrill|2022}} ==== Lack of transparency ==== {{See also|Explainable AI|Algorithmic transparency|Right to explanation}} Many AI systems are so complex that their designers cannot explain how they reach their decisions.{{Sfnp|Sample|2017}} Particularly with [[deep neural networks]], in which there are a large amount of non-[[linear]] relationships between inputs and outputs. But some popular explainability techniques exist.<ref>{{Cite web |date=16 June 2023 |title=Black Box AI |url=https://www.techopedia.com/definition/34940/black-box-ai |access-date=5 October 2024 |archive-date=15 June 2024 |archive-url=https://web.archive.org/web/20240615100800/https://www.techopedia.com/definition/34940/black-box-ai |url-status=live }}</ref> It is impossible to be certain that a program is operating correctly if no one knows how exactly it works. There have been many cases where a machine learning program passed rigorous tests, but nevertheless learned something different than what the programmers intended. For example, a system that could identify skin diseases better than medical professionals was found to actually have a strong tendency to classify images with a [[ruler]] as "cancerous", because pictures of malignancies typically include a ruler to show the scale.{{Sfnp|Christian|2020|p=110}} Another machine learning system designed to help effectively allocate medical resources was found to classify patients with asthma as being at "low risk" of dying from pneumonia. Having asthma is actually a severe risk factor, but since the patients having asthma would usually get much more medical care, they were relatively unlikely to die according to the training data. The correlation between asthma and low risk of dying from pneumonia was real, but misleading.{{Sfnp|Christian|2020|pp=88–91}} People who have been harmed by an algorithm's decision have a right to an explanation.<ref>{{Harvtxt|Christian|2020|p=83}}; {{Harvtxt|Russell|Norvig|2021|p=997}}</ref> Doctors, for example, are expected to clearly and completely explain to their colleagues the reasoning behind any decision they make. Early drafts of the European Union's [[General Data Protection Regulation]] in 2016 included an explicit statement that this right exists.{{Efn|When the law was passed in 2018, it still contained a form of this provision.}} Industry experts noted that this is an unsolved problem with no solution in sight. Regulators argued that nevertheless the harm is real: if the problem has no solution, the tools should not be used.{{Sfnp|Christian|2020|p=91}} [[DARPA]] established the [[Explainable Artificial Intelligence|XAI]] ("Explainable Artificial Intelligence") program in 2014 to try to solve these problems.{{Sfnp|Christian|2020|p=83}} Several approaches aim to address the transparency problem. SHAP enables to visualise the contribution of each feature to the output.{{Sfnp|Verma|2021}} LIME can locally approximate a model's outputs with a simpler, interpretable model.{{Sfnp|Rothman|2020}} [[Multitask learning]] provides a large number of outputs in addition to the target classification. These other outputs can help developers deduce what the network has learned.{{Sfnp|Christian|2020|pp=105–108}} [[Deconvolution]], [[DeepDream]] and other [[generative AI|generative]] methods can allow developers to see what different layers of a deep network for computer vision have learned, and produce output that can suggest what the network is learning.{{Sfnp|Christian|2020|pp=108–112}} For [[generative pre-trained transformer]]s, [[Anthropic]] developed a technique based on [[dictionary learning]] that associates patterns of neuron activations with human-understandable concepts.<ref>{{Cite web |last=Ropek |first=Lucas |date=2024-05-21 |title=New Anthropic Research Sheds Light on AI's 'Black Box' |url=https://gizmodo.com/new-anthropic-research-sheds-light-on-ais-black-box-1851491333 |access-date=2024-05-23 |website=Gizmodo |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170309/https://gizmodo.com/new-anthropic-research-sheds-light-on-ais-black-box-1851491333 |url-status=live }}</ref> ==== Bad actors and weaponized AI ==== {{Main|Lethal autonomous weapon|Artificial intelligence arms race|AI safety}} Artificial intelligence provides a number of tools that are useful to [[bad actor]]s, such as [[authoritarian|authoritarian governments]], [[terrorist]]s, [[criminals]] or [[rogue states]]. A lethal autonomous weapon is a machine that locates, selects and engages human targets without human supervision.{{Efn|This is the [[United Nations]]' definition, and includes things like [[land mines]] as well.{{Sfnp|Russell|Norvig|2021|p=989}}}} Widely available AI tools can be used by bad actors to develop inexpensive autonomous weapons and, if produced at scale, they are potentially [[weapons of mass destruction]].{{Sfnp|Russell|Norvig|2021|pp=987–990}} Even when used in conventional warfare, they currently cannot reliably choose targets and could potentially [[murder|kill an innocent person]].{{Sfnp|Russell|Norvig|2021|pp=987–990}} In 2014, 30 nations (including China) supported a ban on autonomous weapons under the [[United Nations]]' [[Convention on Certain Conventional Weapons]], however the [[United States]] and others disagreed.{{Sfnp|Russell|Norvig|2021|p=988}} By 2015, over fifty countries were reported to be researching battlefield robots.<ref>{{Harvtxt|Robitzski|2018}}; {{Harvtxt|Sainato|2015}}</ref> AI tools make it easier for [[Authoritarian|authoritarian governments]] to efficiently control their citizens in several ways. [[Facial recognition system|Face]] and [[Speaker recognition|voice recognition]] allow widespread [[surveillance]]. [[Machine learning]], operating this data, can [[classifier (machine learning)|classify]] potential enemies of the state and prevent them from hiding. [[Recommendation systems]] can precisely target [[propaganda]] and [[misinformation]] for maximum effect. [[Deepfakes]] and [[generative AI]] aid in producing misinformation. Advanced AI can make authoritarian [[technocracy|centralized decision making]] more competitive than liberal and decentralized systems such as [[market (economics)|market]]s. It lowers the cost and difficulty of [[digital warfare]] and [[spyware|advanced spyware]].{{Sfnp|Harari|2018}} All these technologies have been available since 2020 or earlier—AI [[facial recognition system]]s are already being used for [[mass surveillance]] in China.<ref>{{Cite news |last1=Buckley |first1=Chris |last2=Mozur |first2=Paul |date=22 May 2019 |title=How China Uses High-Tech Surveillance to Subdue Minorities |url=https://www.nytimes.com/2019/05/22/world/asia/china-surveillance-xinjiang.html |work=The New York Times |access-date=2 July 2019 |archive-date=25 November 2019 |archive-url=https://web.archive.org/web/20191125180459/https://www.nytimes.com/2019/05/22/world/asia/china-surveillance-xinjiang.html |url-status=live }}</ref><ref>{{Cite web |date=3 May 2019 |title=Security lapse exposed a Chinese smart city surveillance system |url=https://techcrunch.com/2019/05/03/china-smart-city-exposed |url-status=live |archive-url=https://web.archive.org/web/20210307203740/https://consent.yahoo.com/v2/collectConsent?sessionId=3_cc-session_c8562b93-9863-4915-8523-6c7b930a3efc |archive-date=7 March 2021 |access-date=14 September 2020}}</ref> There many other ways that AI is expected to help bad actors, some of which can not be foreseen. For example, machine-learning AI is able to design tens of thousands of toxic molecules in a matter of hours.{{Sfnp|Urbina|Lentzos|Invernizzi|Ekins|2022}} ==== Technological unemployment ==== {{Main|Workplace impact of artificial intelligence|Technological unemployment}} Economists have frequently highlighted the risks of redundancies from AI, and speculated about unemployment if there is no adequate social policy for full employment.<ref name="E">E. McGaughey, 'Will Robots Automate Your Job Away? Full Employment, Basic Income, and Economic Democracy' (2022), [https://academic.oup.com/ilj/article/51/3/511/6321008 51(3) Industrial Law Journal 511–559]. {{Webarchive|url=https://web.archive.org/web/20230527163045/https://academic.oup.com/ilj/article/51/3/511/6321008|date=27 May 2023}}.</ref> <!-- TOPIC: ESTIMATES OF THE AMOUNT OF UNEMPLOYMENT --> In the past, technology has tended to increase rather than reduce total employment, but economists acknowledge that "we're in uncharted territory" with AI.<ref>{{Harvtxt|Ford|Colvin|2015}};{{Harvtxt|McGaughey|2022}}</ref> A survey of economists showed disagreement about whether the increasing use of robots and AI will cause a substantial increase in long-term [[unemployment]], but they generally agree that it could be a net benefit if [[productivity]] gains are [[Redistribution of income and wealth|redistributed]].{{Sfnp|IGM Chicago|2017}} Risk estimates vary; for example, in the 2010s, Michael Osborne and [[Carl Benedikt Frey]] estimated 47% of U.S. jobs are at "high risk" of potential automation, while an OECD report classified only 9% of U.S. jobs as "high risk".{{Efn|See table 4; 9% is both the OECD average and the U.S. average.{{Sfnp|Arntz|Gregory|Zierahn|2016|p=33}}}}<ref>{{Harvtxt|Lohr|2017}}; {{Harvtxt|Frey|Osborne|2017}}; {{Harvtxt|Arntz|Gregory|Zierahn|2016|p=33}}</ref> The methodology of speculating about future employment levels has been criticised as lacking evidential foundation, and for implying that technology, rather than social policy, creates unemployment, as opposed to redundancies.<ref name="E"/> In April 2023, it was reported that 70% of the jobs for Chinese video game illustrators had been eliminated by generative artificial intelligence.<ref>{{Cite web |last=Zhou |first=Viola |date=2023-04-11 |title=AI is already taking video game illustrators' jobs in China |url=https://restofworld.org/2023/ai-image-china-video-game-layoffs |access-date=2023-08-17 |website=Rest of World |archive-date=21 February 2024 |archive-url=https://web.archive.org/web/20240221131748/https://restofworld.org/2023/ai-image-china-video-game-layoffs/ |url-status=live }}</ref><ref>{{Cite web |last=Carter |first=Justin |date=2023-04-11 |title=China's game art industry reportedly decimated by growing AI use |url=https://www.gamedeveloper.com/art/china-s-game-art-industry-reportedly-decimated-ai-art-use |access-date=2023-08-17 |website=Game Developer |archive-date=17 August 2023 |archive-url=https://web.archive.org/web/20230817010519/https://www.gamedeveloper.com/art/china-s-game-art-industry-reportedly-decimated-ai-art-use |url-status=live }}</ref> <!-- TOPIC: WHICH JOBS ARE AT RISK? --> Unlike previous waves of automation, many middle-class jobs may be eliminated by artificial intelligence; ''[[The Economist]]'' stated in 2015 that "the worry that AI could do to white-collar jobs what steam power did to blue-collar ones during the Industrial Revolution" is "worth taking seriously".{{Sfnp|Morgenstern|2015}} Jobs at extreme risk range from [[paralegal]]s to fast food cooks, while job demand is likely to increase for care-related professions ranging from personal healthcare to the clergy.<ref>{{Harvtxt|Mahdawi|2017}}; {{Harvtxt|Thompson|2014}}</ref> From the early days of the development of artificial intelligence, there have been arguments, for example, those put forward by [[Joseph Weizenbaum]], about whether tasks that can be done by computers actually should be done by them, given the difference between computers and humans, and between quantitative calculation and qualitative, value-based judgement.<ref>{{Cite news |last=Tarnoff |first=Ben |date=4 August 2023 |title=Lessons from Eliza |work=[[The Guardian Weekly]] |pages=34–39}}</ref> ==== Existential risk ==== {{Main|Existential risk from artificial intelligence}} It has been argued AI will become so powerful that humanity may irreversibly lose control of it. This could, as physicist [[Stephen Hawking]] stated, "[[Global catastrophic risk|spell the end of the human race]]".{{Sfnp|Cellan-Jones|2014}} This scenario has been common in science fiction, when a computer or robot suddenly develops a human-like "self-awareness" (or "sentience" or "consciousness") and becomes a malevolent character.{{Efn|Sometimes called a "[[robopocalypse]]"{{Sfn|Russell|Norvig|2021|p=1001}}}} These sci-fi scenarios are misleading in several ways. First, AI does not require human-like [[sentience]] to be an existential risk. Modern AI programs are given specific goals and use learning and intelligence to achieve them. Philosopher [[Nick Bostrom]] argued that if one gives ''almost any'' goal to a sufficiently powerful AI, it may choose to destroy humanity to achieve it (he used the example of a [[Instrumental convergence#Paperclip maximizer|paperclip factory manager]]).{{Sfnp|Bostrom|2014}} [[Stuart J. Russell|Stuart Russell]] gives the example of household robot that tries to find a way to kill its owner to prevent it from being unplugged, reasoning that "you can't fetch the coffee if you're dead."{{Sfnp|Russell|2019}} In order to be safe for humanity, a [[superintelligence]] would have to be genuinely [[AI alignment|aligned]] with humanity's morality and values so that it is "fundamentally on our side".<ref>{{Harvtxt|Bostrom|2014}}; {{Harvtxt|Müller|Bostrom|2014}}; {{Harvtxt|Bostrom|2015}}.</ref> Second, [[Yuval Noah Harari]] argues that AI does not require a robot body or physical control to pose an existential risk. The essential parts of civilization are not physical. Things like [[ideologies]], [[law]], [[government]], [[money]] and the [[economy]] are built on [[language]]; they exist because there are stories that billions of people believe. The current prevalence of [[misinformation]] suggests that an AI could use language to convince people to believe anything, even to take actions that are destructive.{{Sfnp|Harari|2023}} <!-- Warnings of existential risk --> The opinions amongst experts and industry insiders are mixed, with sizable fractions both concerned and unconcerned by risk from eventual superintelligent AI.{{Sfnp|Müller|Bostrom|2014}} Personalities such as [[Stephen Hawking]], [[Bill Gates]], and [[Elon Musk]],<ref>Leaders' concerns about the existential risks of AI around 2015: {{Harvtxt|Rawlinson|2015}}, {{Harvtxt|Holley|2015}}, {{Harvtxt|Gibbs|2014}}, {{Harvtxt|Sainato|2015}}</ref> as well as AI pioneers such as [[Yoshua Bengio]], [[Stuart J. Russell|Stuart Russell]], [[Demis Hassabis]], and [[Sam Altman]], have expressed concerns about existential risk from AI. In May 2023, [[Geoffrey Hinton]] announced his resignation from Google in order to be able to "freely speak out about the risks of AI" without "considering how this impacts Google".<ref>{{Cite news |date=25 March 2023 |title="Godfather of artificial intelligence" talks impact and potential of new AI |url=https://www.cbsnews.com/video/godfather-of-artificial-intelligence-talks-impact-and-potential-of-new-ai |url-status=live |archive-url=https://web.archive.org/web/20230328225221/https://www.cbsnews.com/video/godfather-of-artificial-intelligence-talks-impact-and-potential-of-new-ai |archive-date=28 March 2023 |access-date=2023-03-28 |work=CBS News}}</ref> He notably mentioned risks of an [[AI takeover]],<ref>{{Cite news |last=Pittis |first=Don |date=May 4, 2023 |title=Canadian artificial intelligence leader Geoffrey Hinton piles on fears of computer takeover |url=https://www.cbc.ca/news/business/ai-doom-column-don-pittis-1.6829302 |work=CBC |access-date=5 October 2024 |archive-date=7 July 2024 |archive-url=https://web.archive.org/web/20240707032135/https://www.cbc.ca/news/business/ai-doom-column-don-pittis-1.6829302 |url-status=live }}</ref> and stressed that in order to avoid the worst outcomes, establishing safety guidelines will require cooperation among those competing in use of AI.<ref>{{Cite web |date=2024-06-14 |title='50–50 chance' that AI outsmarts humanity, Geoffrey Hinton says |url=https://www.bnnbloomberg.ca/50-50-chance-that-ai-outsmarts-humanity-geoffrey-hinton-says-1.2085394 |access-date=2024-07-06 |website=Bloomberg BNN |archive-date=14 June 2024 |archive-url=https://web.archive.org/web/20240614144506/https://www.bnnbloomberg.ca/50-50-chance-that-ai-outsmarts-humanity-geoffrey-hinton-says-1.2085394 |url-status=live }}</ref> In 2023, many leading AI experts endorsed [[Statement on AI risk of extinction|the joint statement]] that "Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war".{{Sfnp|Valance|2023}} <!-- Arguments against existential risk --> Some other researchers were more optimistic. AI pioneer [[Jürgen Schmidhuber]] did not sign the joint statement, emphasising that in 95% of all cases, AI research is about making "human lives longer and healthier and easier."<ref>{{Cite news |last=Taylor |first=Josh |date=7 May 2023 |title=Rise of artificial intelligence is inevitable but should not be feared, 'father of AI' says |url=https://www.theguardian.com/technology/2023/may/07/rise-of-artificial-intelligence-is-inevitable-but-should-not-be-feared-father-of-ai-says |access-date=26 May 2023 |work=The Guardian |archive-date=23 October 2023 |archive-url=https://web.archive.org/web/20231023061228/https://www.theguardian.com/technology/2023/may/07/rise-of-artificial-intelligence-is-inevitable-but-should-not-be-feared-father-of-ai-says |url-status=live }}</ref> While the tools that are now being used to improve lives can also be used by bad actors, "they can also be used against the bad actors."<ref>{{Cite news |last=Colton |first=Emma |date=7 May 2023 |title='Father of AI' says tech fears misplaced: 'You cannot stop it' |url=https://www.foxnews.com/tech/father-ai-jurgen-schmidhuber-says-tech-fears-misplaced-cannot-stop |access-date=26 May 2023 |work=Fox News |archive-date=26 May 2023 |archive-url=https://web.archive.org/web/20230526162642/https://www.foxnews.com/tech/father-ai-jurgen-schmidhuber-says-tech-fears-misplaced-cannot-stop |url-status=live }}</ref><ref>{{Cite news |last=Jones |first=Hessie |date=23 May 2023 |title=Juergen Schmidhuber, Renowned 'Father Of Modern AI,' Says His Life's Work Won't Lead To Dystopia |url=https://www.forbes.com/sites/hessiejones/2023/05/23/juergen-schmidhuber-renowned-father-of-modern-ai-says-his-lifes-work-wont-lead-to-dystopia |access-date=26 May 2023 |work=Forbes |archive-date=26 May 2023 |archive-url=https://web.archive.org/web/20230526163102/https://www.forbes.com/sites/hessiejones/2023/05/23/juergen-schmidhuber-renowned-father-of-modern-ai-says-his-lifes-work-wont-lead-to-dystopia/ |url-status=live }}</ref> [[Andrew Ng]] also argued that "it's a mistake to fall for the doomsday hype on AI—and that regulators who do will only benefit vested interests."<ref>{{Cite news |last=McMorrow |first=Ryan |date=19 Dec 2023 |title=Andrew Ng: 'Do we think the world is better off with more or less intelligence?' |url=https://www.ft.com/content/2dc07f9e-d2a9-4d98-b746-b051f9352be3 |access-date=30 Dec 2023 |work=Financial Times |archive-date=25 January 2024 |archive-url=https://web.archive.org/web/20240125014121/https://www.ft.com/content/2dc07f9e-d2a9-4d98-b746-b051f9352be3 |url-status=live }}</ref> [[Yann LeCun]] "scoffs at his peers' dystopian scenarios of supercharged misinformation and even, eventually, human extinction."<ref>{{Cite magazine |last=Levy |first=Steven |date=22 Dec 2023 |title=How Not to Be Stupid About AI, With Yann LeCun |url=https://www.wired.com/story/artificial-intelligence-meta-yann-lecun-interview |access-date=30 Dec 2023 |magazine=Wired |archive-date=28 December 2023 |archive-url=https://web.archive.org/web/20231228152443/https://www.wired.com/story/artificial-intelligence-meta-yann-lecun-interview/ |url-status=live }}</ref> In the early 2010s, experts argued that the risks are too distant in the future to warrant research or that humans will be valuable from the perspective of a superintelligent machine.<ref>Arguments that AI is not an imminent risk: {{Harvtxt|Brooks|2014}}, {{Harvtxt|Geist|2015}}, {{Harvtxt|Madrigal|2015}}, {{Harvtxt|Lee|2014}}</ref> However, after 2016, the study of current and future risks and possible solutions became a serious area of research.{{Sfnp|Christian|2020|pp=67, 73}} === Ethical machines and alignment === {{Main|Machine ethics|AI safety|Friendly artificial intelligence|Artificial moral agents|Human Compatible}} Friendly AI are machines that have been designed from the beginning to minimize risks and to make choices that benefit humans. [[Eliezer Yudkowsky]], who coined the term, argues that developing friendly AI should be a higher research priority: it may require a large investment and it must be completed before AI becomes an existential risk.{{Sfnp|Yudkowsky|2008}} Machines with intelligence have the potential to use their intelligence to make ethical decisions. The field of machine ethics provides machines with ethical principles and procedures for resolving ethical dilemmas.{{Sfnp|Anderson|Anderson|2011}} The field of machine ethics is also called computational morality,{{Sfnp|Anderson|Anderson|2011}} and was founded at an [[AAAI]] symposium in 2005.{{Sfnp|AAAI|2014}} Other approaches include [[Wendell Wallach]]'s "artificial moral agents"{{Sfnp|Wallach|2010}} and [[Stuart J. Russell]]'s [[Human Compatible#Russell's three principles|three principles]] for developing provably beneficial machines.{{Sfnp|Russell|2019|p=173}} === Open source === Active organizations in the AI open-source community include [[Hugging Face]],<ref>{{Cite web |last1=Stewart |first1=Ashley |last2=Melton |first2=Monica |title=Hugging Face CEO says he's focused on building a 'sustainable model' for the $4.5 billion open-source-AI startup |url=https://www.businessinsider.com/hugging-face-open-source-ai-approach-2023-12 |access-date=2024-04-14 |website=Business Insider |archive-date=25 September 2024 |archive-url=https://web.archive.org/web/20240925013220/https://www.businessinsider.com/hugging-face-open-source-ai-approach-2023-12 |url-status=live }}</ref> [[Google]],<ref>{{Cite web |last=Wiggers |first=Kyle |date=2024-04-09 |title=Google open sources tools to support AI model development |url=https://techcrunch.com/2024/04/09/google-open-sources-tools-to-support-ai-model-development |access-date=2024-04-14 |website=TechCrunch |archive-date=10 September 2024 |archive-url=https://web.archive.org/web/20240910112401/https://techcrunch.com/2024/04/09/google-open-sources-tools-to-support-ai-model-development/ |url-status=live }}</ref> [[EleutherAI]] and [[Meta Platforms|Meta]].<ref>{{Cite web |last=Heaven |first=Will Douglas |date=May 12, 2023 |title=The open-source AI boom is built on Big Tech's handouts. How long will it last? |url=https://www.technologyreview.com/2023/05/12/1072950/open-source-ai-google-openai-eleuther-meta |access-date=2024-04-14 |website=MIT Technology Review}}</ref> Various AI models, such as [[LLaMA|Llama 2]], [[Mistral AI|Mistral]] or [[Stable Diffusion]], have been made open-weight,<ref>{{Cite news |last=Brodsky |first=Sascha |date=December 19, 2023 |title=Mistral AI's New Language Model Aims for Open Source Supremacy |url=https://aibusiness.com/nlp/mistral-ai-s-new-language-model-aims-for-open-source-supremacy |work=AI Business |access-date=5 October 2024 |archive-date=5 September 2024 |archive-url=https://web.archive.org/web/20240905212607/https://aibusiness.com/nlp/mistral-ai-s-new-language-model-aims-for-open-source-supremacy |url-status=live }}</ref><ref>{{Cite web |last=Edwards |first=Benj |date=2024-02-22 |title=Stability announces Stable Diffusion 3, a next-gen AI image generator |url=https://arstechnica.com/information-technology/2024/02/stability-announces-stable-diffusion-3-a-next-gen-ai-image-generator |access-date=2024-04-14 |website=Ars Technica |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170201/https://arstechnica.com/information-technology/2024/02/stability-announces-stable-diffusion-3-a-next-gen-ai-image-generator/ |url-status=live }}</ref> meaning that their architecture and trained parameters (the "weights") are publicly available. Open-weight models can be freely [[Fine-tuning (deep learning)|fine-tuned]], which allows companies to specialize them with their own data and for their own use-case.<ref>{{Cite news |last=Marshall |first=Matt |date=January 29, 2024 |title=How enterprises are using open source LLMs: 16 examples |url=https://venturebeat.com/ai/how-enterprises-are-using-open-source-llms-16-examples |work=VentureBeat |access-date=5 October 2024 |archive-date=26 September 2024 |archive-url=https://web.archive.org/web/20240926171131/https://venturebeat.com/ai/how-enterprises-are-using-open-source-llms-16-examples/ |url-status=live }}</ref> Open-weight models are useful for research and innovation but can also be misused. Since they can be fine-tuned, any built-in security measure, such as objecting to harmful requests, can be trained away until it becomes ineffective. Some researchers warn that future AI models may develop dangerous capabilities (such as the potential to drastically facilitate [[bioterrorism]]) and that once released on the Internet, they cannot be deleted everywhere if needed. They recommend pre-release audits and cost-benefit analyses.<ref>{{Cite web |last=Piper |first=Kelsey |date=2024-02-02 |title=Should we make our most powerful AI models open source to all? |url=https://www.vox.com/future-perfect/2024/2/2/24058484/open-source-artificial-intelligence-ai-risk-meta-llama-2-chatgpt-openai-deepfake |access-date=2024-04-14 |website=Vox |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170204/https://www.vox.com/future-perfect/2024/2/2/24058484/open-source-artificial-intelligence-ai-risk-meta-llama-2-chatgpt-openai-deepfake |url-status=live }}</ref> === Frameworks === Artificial Intelligence projects can be guided by ethical considerations during the design, development, and implementation of an AI system. An AI framework such as the Care and Act Framework, developed by the [[Alan Turing Institute]] and based on the SUM values, outlines four main ethical dimensions, defined as follows:<ref>{{Cite web |author=Alan Turing Institute |date=2019 |title=Understanding artificial intelligence ethics and safety |url=https://www.turing.ac.uk/sites/default/files/2019-06/understanding_artificial_intelligence_ethics_and_safety.pdf |access-date=5 October 2024 |archive-date=11 September 2024 |archive-url=https://web.archive.org/web/20240911131935/https://www.turing.ac.uk/sites/default/files/2019-06/understanding_artificial_intelligence_ethics_and_safety.pdf |url-status=live }}</ref><ref>{{Cite web |author=Alan Turing Institute |date=2023 |title=AI Ethics and Governance in Practice |url=https://www.turing.ac.uk/sites/default/files/2023-12/aieg-ati-ai-ethics-an-intro_1.pdf |access-date=5 October 2024 |archive-date=11 September 2024 |archive-url=https://web.archive.org/web/20240911125504/https://www.turing.ac.uk/sites/default/files/2023-12/aieg-ati-ai-ethics-an-intro_1.pdf |url-status=live }}</ref> * '''Respect''' the dignity of individual people * '''Connect''' with other people sincerely, openly, and inclusively * '''Care''' for the wellbeing of everyone * '''Protect''' social values, justice, and the public interest Other developments in ethical frameworks include those decided upon during the [[Asilomar Conference on Beneficial AI|Asilomar Conference]], the Montreal Declaration for Responsible AI, and the IEEE's Ethics of Autonomous Systems initiative, among others;<ref>{{Cite journal |last1=Floridi |first1=Luciano |last2=Cowls |first2=Josh |date=2019-06-23 |title=A Unified Framework of Five Principles for AI in Society |url=https://hdsr.mitpress.mit.edu/pub/l0jsh9d1 |journal=Harvard Data Science Review |volume=1 |issue=1 |doi=10.1162/99608f92.8cd550d1 |s2cid=198775713 |doi-access=free |archive-date=7 August 2019 |access-date=5 December 2023 |archive-url=https://archive.today/20190807202909/https://hdsr.mitpress.mit.edu/pub/l0jsh9d1 |url-status=live }}</ref> however, these principles are not without criticism, especially regards to the people chosen to contribute to these frameworks.<ref>{{Cite journal |last1=Buruk |first1=Banu |last2=Ekmekci |first2=Perihan Elif |last3=Arda |first3=Berna |date=2020-09-01 |title=A critical perspective on guidelines for responsible and trustworthy artificial intelligence |url=https://doi.org/10.1007/s11019-020-09948-1 |journal=Medicine, Health Care and Philosophy |volume=23 |issue=3 |pages=387–399 |doi=10.1007/s11019-020-09948-1 |issn=1572-8633 |pmid=32236794 |s2cid=214766800 |access-date=5 October 2024 |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170206/https://link.springer.com/article/10.1007/s11019-020-09948-1 |url-status=live }}</ref> Promotion of the wellbeing of the people and communities that these technologies affect requires consideration of the social and ethical implications at all stages of AI system design, development and implementation, and collaboration between job roles such as data scientists, product managers, data engineers, domain experts, and delivery managers.<ref>{{Cite journal |last1=Kamila |first1=Manoj Kumar |last2=Jasrotia |first2=Sahil Singh |date=2023-01-01 |title=Ethical issues in the development of artificial intelligence: recognizing the risks |url=https://doi.org/10.1108/IJOES-05-2023-0107 |journal=International Journal of Ethics and Systems |pages=45–63 |volume=41 |issue=ahead-of-print |doi=10.1108/IJOES-05-2023-0107 |issn=2514-9369 |s2cid=259614124 |access-date=5 October 2024 |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170207/https://www.emerald.com/insight/content/doi/10.1108/IJOES-05-2023-0107/full/html |url-status=live }}</ref> The [[AI Safety Institute (United Kingdom)|UK AI Safety Institute]] released in 2024 a testing toolset called 'Inspect' for AI safety evaluations available under a MIT open-source licence which is freely available on GitHub and can be improved with third-party packages. It can be used to evaluate AI models in a range of areas including core knowledge, ability to reason, and autonomous capabilities.<ref>{{Cite web |date=10 May 2024 |title=AI Safety Institute releases new AI safety evaluations platform |url=https://www.gov.uk/government/news/ai-safety-institute-releases-new-ai-safety-evaluations-platform |access-date=14 May 2024 |publisher=UK Government |archive-date=5 October 2024 |archive-url=https://web.archive.org/web/20241005170207/https://www.gov.uk/government/news/ai-safety-institute-releases-new-ai-safety-evaluations-platform |url-status=live }}</ref> === Regulation === {{Main|Regulation of artificial intelligence|Regulation of algorithms|AI safety}} [[File:Vice President Harris at the group photo of the 2023 AI Safety Summit.jpg|upright=1.2|thumb|alt=AI Safety Summit|The first global [[AI Safety Summit]] was held in the United Kingdom in November 2023 with a declaration calling for international cooperation.]] The regulation of artificial intelligence is the development of public sector policies and laws for promoting and regulating AI; it is therefore related to the broader regulation of algorithms.<ref>Regulation of AI to mitigate risks: {{Harvtxt|Berryhill|Heang|Clogher|McBride|2019}}, {{Harvtxt|Barfield|Pagallo|2018}}, {{Harvtxt|Iphofen|Kritikos|2019}}, {{Harvtxt|Wirtz|Weyerer|Geyer|2018}}, {{Harvtxt|Buiten|2019}}</ref> The regulatory and policy landscape for AI is an emerging issue in jurisdictions globally.{{Sfnp|Law Library of Congress (U.S.). Global Legal Research Directorate|2019}} According to AI Index at [[Stanford]], the annual number of AI-related laws passed in the 127 survey countries jumped from one passed in 2016 to 37 passed in 2022 alone.{{Sfnp|Vincent|2023}}{{Sfnp|Stanford University|2023}} Between 2016 and 2020, more than 30 countries adopted dedicated strategies for AI.{{Sfnp|UNESCO|2021}} Most EU member states had released national AI strategies, as had Canada, China, India, Japan, Mauritius, the Russian Federation, Saudi Arabia, United Arab Emirates, U.S., and Vietnam. Others were in the process of elaborating their own AI strategy, including Bangladesh, Malaysia and Tunisia.{{Sfnp|UNESCO|2021}} The [[Global Partnership on Artificial Intelligence]] was launched in June 2020, stating a need for AI to be developed in accordance with human rights and democratic values, to ensure public confidence and trust in the technology.{{Sfnp|UNESCO|2021}} [[Henry Kissinger]], [[Eric Schmidt]], and [[Daniel Huttenlocher]] published a joint statement in November 2021 calling for a government commission to regulate AI.{{Sfnp|Kissinger|2021}} In 2023, OpenAI leaders published recommendations for the governance of superintelligence, which they believe may happen in less than 10 years.{{Sfnp|Altman|Brockman|Sutskever |2023}} In 2023, the United Nations also launched an advisory body to provide recommendations on AI governance; the body comprises technology company executives, governments officials and academics.<ref>{{Cite web |last=VOA News |date=October 25, 2023 |title=UN Announces Advisory Body on Artificial Intelligence |url=https://www.voanews.com/a/un-announces-advisory-body-on-artificial-intelligence-/7328732.html |access-date=5 October 2024 |archive-date=18 September 2024 |archive-url=https://web.archive.org/web/20240918071530/https://www.voanews.com/a/un-announces-advisory-body-on-artificial-intelligence-/7328732.html |url-status=live }}</ref> In 2024, the [[Council of Europe]] created the first international legally binding treaty on AI, called the "[[Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law]]". It was adopted by the European Union, the United States, the United Kingdom, and other signatories.<ref>{{Cite web |date=5 September 2024 |title=Council of Europe opens first ever global treaty on AI for signature |url=https://www.coe.int/en/web/portal/-/council-of-europe-opens-first-ever-global-treaty-on-ai-for-signature |access-date=2024-09-17 |website=Council of Europe |archive-date=17 September 2024 |archive-url=https://web.archive.org/web/20240917001330/https://www.coe.int/en/web/portal/-/council-of-europe-opens-first-ever-global-treaty-on-ai-for-signature |url-status=live }}</ref> In a 2022 [[Ipsos]] survey, attitudes towards AI varied greatly by country; 78% of Chinese citizens, but only 35% of Americans, agreed that "products and services using AI have more benefits than drawbacks".{{Sfnp|Vincent|2023}} A 2023 [[Reuters]]/Ipsos poll found that 61% of Americans agree, and 22% disagree, that AI poses risks to humanity.{{Sfnp|Edwards|2023}} In a 2023 [[Fox News]] poll, 35% of Americans thought it "very important", and an additional 41% thought it "somewhat important", for the federal government to regulate AI, versus 13% responding "not very important" and 8% responding "not at all important".{{Sfnp|Kasperowicz|2023}}{{Sfnp|Fox News|2023}} In November 2023, the first global [[AI Safety Summit]] was held in [[Bletchley Park]] in the UK to discuss the near and far term risks of AI and the possibility of mandatory and voluntary regulatory frameworks.<ref>{{Cite news |last=Milmo |first=Dan |date=3 November 2023 |title=Hope or Horror? The great AI debate dividing its pioneers |work=[[The Guardian Weekly]] |pages=10–12}}</ref> 28 countries including the United States, China, and the European Union issued a declaration at the start of the summit, calling for international co-operation to manage the challenges and risks of artificial intelligence.<ref>{{Cite web |date=1 November 2023 |title=The Bletchley Declaration by Countries Attending the AI Safety Summit, 1–2 November 2023 |url=https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023 |archive-url=https://web.archive.org/web/20231101123904/https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023 |archive-date=1 November 2023 |access-date=2 November 2023 |website=GOV.UK}}</ref><ref>{{Cite press release |title=Countries agree to safe and responsible development of frontier AI in landmark Bletchley Declaration |url=https://www.gov.uk/government/news/countries-agree-to-safe-and-responsible-development-of-frontier-ai-in-landmark-bletchley-declaration |access-date=1 November 2023 |url-status=live |archive-url=https://web.archive.org/web/20231101115016/https://www.gov.uk/government/news/countries-agree-to-safe-and-responsible-development-of-frontier-ai-in-landmark-bletchley-declaration |archive-date=1 November 2023 |website=GOV.UK}}</ref> In May 2024 at the [[AI Seoul Summit]], 16 global AI tech companies agreed to safety commitments on the development of AI.<ref>{{Cite web |date=21 May 2024 |title=Second global AI summit secures safety commitments from companies |url=https://www.reuters.com/technology/global-ai-summit-seoul-aims-forge-new-regulatory-agreements-2024-05-21 |access-date=23 May 2024 |publisher=Reuters}}</ref><ref>{{Cite web |date=21 May 2024 |title=Frontier AI Safety Commitments, AI Seoul Summit 2024 |url=https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024 |archive-url=https://web.archive.org/web/20240523201611/https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024 |archive-date=23 May 2024 |access-date=23 May 2024 |publisher=gov.uk}}</ref>
Edit summary
(Briefly describe your changes)
By publishing changes, you agree to the
Terms of Use
, and you irrevocably agree to release your contribution under the
CC BY-SA 4.0 License
and the
GFDL
. You agree that a hyperlink or URL is sufficient attribution under the Creative Commons license.
Cancel
Editing help
(opens in new window)